Skill-based Rankings and Skill-Based Matchmaking are Relevant to Brink:
Having followed this discussion and having heard the developers (Splash Damage) talk about achieving a balanced, team-oriented multiplayer experience, I think the idea of an SC2-esque matchmaking system is quite relevant to Brink.
BUT I don’t think we will be seeing it at Launch
However, I doubt that this feature will be implemented in the retail game, because, quite frankly, we haven’t heard anything about it for Brink. The only multiplayer balancing feature that I can recall is the ‘enemy intel mission’ for floundering players (which has been scrapped, to the best of my knowledge).
On a related note, to those of you who doubt that skill-based / ranking systems are relevant to Brink then please read on:
To counter your objections, it would be useful for me to highlight a few videogames that have successfully employed ranking systems and skill-based matchmaking systems (not a comprehensive list):
Games that Employ Skill-Based Matchmaking Systems Effectively:
Starcraft 2
Halo 3
Halo: Reach
As has been previously mentioned, these games employ ‘ELO’ systems. An ELO system ranks players based on their wins and losses, and it assigns them a ranking based on not only the numbers of wins (and losses) but also based upon the quality (strength) of their opponents.
In Halo 3, you saw the indicator of your rating in the form of your ‘skill’ in ranked matches (a number between 1 and 50). The true value of your ELO ranking was hidden from the player in Halo 3.
In Halo:Reach, numerical indications of your ELO rating are hidden from the player, but skill-based matchmaking can be turned on in your preferences. Likewise, there is an ‘Arena’ where teams can compete and will be ranked into Bronze, Silver, Gold, or Onyx leagues (indicative of skill).
Similarly, Starcraft 2 relies on an ELO system to rank players into leagues indicative of their skill level and mastery of the game. i.e. Bronze, Silver, Gold, Platinum, Diamond
An Explanation of the ELO System for Ranking Individuals, Random Groups of Individuals, and Teams:
The way that all of these ELO systems ranks players is as follows:
Set the player’s ranking so that if they were to match up against evenly ranked players, then the likelihood of victory (or defeat) is roughly 50% (i.e. 50% W/L ratio).
Over time, if a player outperforms the ranking as assigned by the system (solely on the basis of wins and losses, not based upon in-game performance), then their ranking (i.e. ‘skill’ as acknowledged by the system) must increase until a 50% win rate is achieved (in this case, by matching the player up against more difficult opponents until their W/L ratio decreased to ~50%).
The cool thing about the ELO system is that it can be applied to individuals, random groups of individuals, or teams. The system simply applies a ranking to each unique group of players. In the case of Starcraft 2, for instance, you could even be assigned a ranking for your performance with ‘Random’ teammates.
Behind the scenes, a database stores (and updates) your score in the ELO system based on your performance for each unique group of players. The system becomes more ‘confident’ as you accumulate more games played with that same group of unique players. In practice, the system relies on two variables to do this: one variable stores the crude estimate of your skill level; another variable represents the variance that it is willing to apply based upon the number of games that you have played. This second variable decreases in magnitude as you play more games and the system becomes more ‘confident’ with your overall ranking. Accordingly, the ranking for a consistent group of players (i.e. an organized team that always plays together) would be more ‘accurate’ (and ‘confident’) than for a random group of players. However, it is still possible to rank the random group of players.
i.e. ‘Skill’ of player = True skill variable vale +/- Variance variable (that converges over time)
ELO and Starcraft 2:
The Starcraft 2 developers (Blizzard) chose to implement an ELO matchmaking system as the primary method of matchmaking, because of the extremely wide variance in skill across the SC2 community. They wanted to preserve a fun gameplay experience for players by ensuring that they didn’t just continually stomp or completely decimate all of their opponents. They reasoned that Starcraft 2 is more fun when you win roughly 50% of the time and play competitive games against evenly matched opponents.
As anyone who is familiar with Starcraft 2 knows, you can be ranked as an individual in 1v1; as a team in 2v2, 3v3, and 4v4; or as a random team player in 2v2, 3v3, or 4v4. The point is: ranking with ELO works even when you, as an individual, are a random player playing on teams of variable teammates.
ELO and the Halo Franchise:
It is worth noting that Halo 3 and Halo:Reach handle skill-based rankings differently.
In Halo 3, the ELO system is based upon your W/L ratio.
i.e. Skill_Rating_new = Skill_Rating_old +/- Bonus_weighted_by_opponent_ranking_and_skill
In Halo:Reach, (for at least one type of ranking) the developers came up with a crude algorithm that both correlates well with the aforementioned skill-based rating but is more geared towards being an effective team-player (as opposed to a successful lone-wolf).
i.e. In-Game Performance Rating = Kills + Assists - Deaths/(~3)
i.e. Skill_new = Skill_old +/- Relative_Performance_Rating_weighted_by_Opponent_Ranking_and_Skill
Both the systems of Halo 3 and Halo:Reach approximate player skill in order to rank and sort players for a more competitive (and satisfying [by their reasoning]) matchmaking experience.
In a game like Halo 3 or Halo:Reach, there may also be a wide variance in skill, but it seems to me that the gameplay experience isn’t compromised as much as it is in the case of Starcraft. There also seems to be some self-selection in the community based on your skill. For instance, I have observed that more casual (and often less ‘skilled’ players) tend to flock to the ‘Rumble Pit’ and ‘Big Team Slayer’ playlists. Likewise, the most hardcore players are more drawn to game modes like objective play, ‘Arena’, or team slayer variants. Furthermore, some of the fun of Halo 3 and Halo:Reach lies in completely dominating lesser opponents (i.e. through better skill, better team-based coordination, using vehicles, or whoring power weapons), and the in-game medal system seems to acknowledge this fact. For example, would a ‘Running Riot’ (15 kill-streak) medal really be relevant, from a game-design perspective, if there was the expectation of a consistent gameplay experience and evenly-matched opponents? Unlikely. Therefore, you can reason that some (if not most) of the fun in Halo 3 and Halo:Reach is in pummeling ‘noobs’. Nonetheless, Halo 3 and Halo:Reach have implemented robust systems for skill-based matchmaking to accommodate competitive players who want to match up against evenly ranked opponents.
The Lack of ELO in the Call of Duty Franchise:
By contrast, I will briefly touch on the Call of Duty franchise. Call of Duty places no emphasis on skill-based or ranked gameplay. There are organized leagues that do rank teams and organize games (i.e. Gamebattles, MLG, etc.), however, these are outside of the scope of Activision and its developers (i.e. InfinityWard, Treyarch, etc.).
As anyone who has played Call of Duty knows, the way that they addresses the issue of skill is by offering players a strong incentive to place themselves at a moderate disadvantage by limiting their kit customization abilities (aka ‘Prestiging’). By doing so, the Call of Duty developers place a premium on the compelling and addictive nature of their gameplay to drive a larger consistent online player population and greater sales, rather than on the addictive properties of highly competitive gameplay (read: skill-based matchmaking for closely matched games) to create a compelling experience for players (to once again drive a larger online player population, more popularity, and better sales).
The end goal of any developer is always the same: to drive sales. Developers simply differ in the approaches they take in catering their gameplay experience to the characteristics of the audience they wish to target.
In Call of Duty, it is the developer’s hope that the addictive offering of Leaderboards (i.e. for tracking total score, score per minute, etc.) and the ‘variable reward’ mechanic of leveling up is a strong enough incentive not only to keep playing but to give you a sense of purpose while you continue to do so. As you progress through the multiplayer experience, players complete challenges and score experience points towards advancing in level. By gaining levels, you unlock new guns, perks, killstreaks, and gadgets that you can deploy in custom class slots to tailor the multiplayer experience to your liking. The structure of this incentive scheme is consistent with that of ‘variable reward’ (although transparent by game design i.e. kill 10 dudes with this gun and get 100 xp ). As anyone who has studied psychology knows, variable reward systems are the most compelling systems for motivating compulsive subject behaviour. As an additional layer of incenting gamer behaviour, there are the oft-mentioned leaderboards. Leaderboards drive hardcore competitive players’ gameplay behaviour to constantly seek to improve their in-game performance. However, as we all know, in-game performance enhancement is not the sole (or necessarily optimal) method to increase leaderboard score and ranking. As such, there can be detrimental side-effects when gamers single-mindedly pursue stat enhancement. Examples of this are common (i.e. not playing an objective, headshot lobbies, etc.). Some collaterals are less obvious. For instance, when an optimal strategy becomes prevalent in a gaming community (as a result of imbalanced game design for instance), then players are punished (in the form of poorer in-game performance) when they deviate from the de facto standard. The combination of the less fun, less successful in-game experience combined with leaderboard performance penalties provides even less incentive for games to be creative with how they approach the game or how they play. You end up with a gameplay experience that is still compelling and addictive, but the lack of variety and stale metagame erodes the fun for all but the most compulsive and hard-core (or utterly casual and oblivious) of your player user-base. The net result: players leave your population of online players; the lifespan of your community wanes; your sales figures don’t realize what “might have been possible”.
Brink’s Development Philosophy for Gameplay Design and Balance:
Splash Damage seems to anticipate having a wide variance of gamer skill in its online Brink multiplayer community.
The developers of Brink are painstakingly using whatever measures possible to not only welcome non-multiplayer gamers into the fold, but to provide them with an experience free of ‘griefers’ and stat-whores who would compromise that experience. Accordingly, Splash Damage has teamplay focused gamplay with a tiered matchmaking system (by level i.e. 1-5, 6-10, etc.) and they have removed leaderboards.
Brink has been designed from the ground up so that single-player is an objective-based mode that mirrors the experience to be had with online multiplayer. To this end, Splash Damage has developed AI that believably mimics or performs at a level consistent with human opponents, and they have designed a communication system and interface that is effective and functions irrespective of computer AI, geographic dialect or language barriers.
Interestingly, Splash Damage seems to focus its offering on delivering a compelling teamplay experience complete with a unique art direction; a vast array of customization options; a novel level navigation mechanic (SMART); and a variable reward xp leveling system to drive compulsive gamer behaviour (up until the level 20 cap, that is).
Likewise, Brink’s developers are trying to develop game balance that does not reward sub-optimal strategies and one that, hopefully, leads to a varied metagame. Ideally, in the event that certain strategies do become metagame-dominant, they will at least have been balanced to coincide with satisfying objective completion and/or promoting teamplay.
So How Is ELO Ranking Relevant to Brink?:
In my opinion, it is from the standpoint of an inexperienced multiplayer gamer, that ELO makes the most sense. My intuition tells me that an inexperienced multiplayer gamer would be more likely to play offline until achieving a high level (say, approaching the level cap of 20) before venturing online.
Under the present system, that gamer would now, for the first time, be facing off against battle-hardened multiplayer gamers who progressed through the level cap against other human opponents. By virtue of this fact, I would expect the multiplayer ‘noob’ to be at a distinct disadvantage.
My supporting assumptions for this are as follows:
- I expect that 'noob’s will be unfamiliar with the more-creative and exploitive strategies of real human players.
- I expect that real human players in parties operating in coordination will be superior to the efforts of a random group of human players.
If the net result of these two assumptions is that an inexperience multiplayer ‘noob’ goes up against real humans for the first time at level cap 20, then they will accumulate loss after loss. This, in turn, will sour their experience with online multiplayer, and they will be reluctant to return as a result.
Therefore, if a ‘noob’ simply loses over and over at a high level cap with the net result that they don’t become multiplayer gamers, then all (or most) of Splash Damage’s efforts to incorporate new multiplayer gamers will have been in vain.
Under this scenario, it would make sense to have an ELO system in place to account for skill variations in those players who arrive at the top matchmaking tier of 16-20 and the level 20 cap through offline play.
If you made it this far, thanks for your time and consideration.